chapter 19 scientiﬁc research based optimisation …...chapter 19 scientiﬁc research based...

Chapter 19Scientific Research Based Optimisationand Geo-information Technologiesfor Integrating Environmental Planningin Disaster Management

Hussain Aziz Saleh and Georges Allaert

Abstract Natural and environmental disasters have profound social, economic,psychological, and demographic effects on the stricken individuals and commu-nities. The literature of disaster management of the 21th Century has pointed outthat there is a missing part in the knowledge, scientific research, and technologi-cal development that can optimise disaster risk reduction. With the improvementof dynamic optimisation and geo-information technologies, it has become veryimportant to determine optimal solutions based on the stability and accuracy of themeasurements that support disaster management and risk reduction. However, a sci-entific approach to the solution of these disasters requires robotic algorithms thatcan provide a degree of functionality for spatial representation and flexibility suit-able for quickly creating optimal solution that account for the uncertainty presentin the changing environment of these disasters. Moreover, the volume of data col-lected for these disasters is growing rapidly, and sophisticated means to optimisethis volume in a consistent, dynamic and economical procedure are essential. Thischapter effectively links wider strategic aims of bringing together innovative waysof thinking based on scientific research, knowledge and technology in many sci-entific disciplines to providing optimal solutions for disaster management and riskreduction. Real-life applications using these disciplines will be presented.

Keywords Disaster management · Risk assessment · Geo-information technol-ogy · Early warning · Artificial intelligence · Dynamic optimization · Environmentalplanning

H.A. Saleh (B)Higher Commission for Scientific Research, P.O. Box 30151, Damascus, Syria;Institute for Sustainable Mobility, Ghent University, Gent, Belgiume-mail: [email protected]; [email protected]; [email protected]

359D. Tang (ed.), Remote Sensing of the Changing Oceans,DOI 10.1007/978-3-642-16541-2_19, C© Springer-Verlag Berlin Heidelberg 2011

360 H.A. Saleh and G. Allaert

1 Introduction

In the past several years, natural disasters which have major impacts in every cornerof the world have dramatically increased. They cannot be prevented and it is notalways possible to completely eliminate their risks, but they can be forecasted attimes, enabling people to properly deal with their consequences. Extensive experi-ence and practice in the past few decades has demonstrated that the damage causedby any disaster can be minimised largely by careful planning, mitigation, and prac-tical actions. For example, with the help of advanced geo-information technology,hurricanes can clearly be seen before they cause devastation and this will enablepreparation and minimisation of expected damage. Early planning has saved lives,but additional planning could have further reduced destruction. Many scientific stud-ies have considered the effects of these disasters, but few have searched for idealsolutions. Scientific research and analysis of hazard data is needed before (riskanalysis, prevention, preparedness), during (emergency aid), and after a disaster(reconstruction) to understand its effect and dimensions. This will help and sup-port determining how best to respond to existing and potential losses, and how toaid effectively with recovery activities. However, risk reduction measures have to beconsidered and evaluated according to several parameters and factors such as social,demographic, and environmental effects, economical cost, available technology, etc.Much more work and research is needed as there are many gaps in our knowledgeand understanding of the changing behaviour of these disasters. In particular, there isa lack of efficient use of geo-information and communication technologies that arepowerful sources for providing accurate information, facilitating communication,and permitting the monitoring of emergency conditions and impacts.

The optimal study for disaster management and risk reduction is based onscientific research, information, knowledge and technology development such aselectronic comparisons using innovative methods that help the decision makers toaccurately understand the relationships between reasons and results, to differen-tiate between the strategic and secondary objectives, and to measure and analysethe gap in performance between the optimal and local models of the disaster.The knowledge methods used in disaster management consist of the varieties ofthe principals and procedures based on scientific research such as trail and error,action and reaction, simulation, modelling, and dynamic optimisation, etc. Also,these methods use performance scales based on e-benchmarking to test and anal-yse the developed techniques for disaster management and mitigation and to definethe improvement and development domains. The information systems presentedby the internet revolution provides and supports databases with all types of infor-mation about disasters and experts anywhere and anytime. While the knowledgesystem is presented with electronic development in the early warning and infor-mation technology through establishing and updating databases, other importantfunctions and innovative methods are essential to achieve results. These functionsand methods are: urgent services, creating information system about all disasters,constructing a site on the net for exchanging information using emails and directcommunication to support the decision makers, doing the continuous improvementsin disaster management, predictions, models, and indicators, etc. This research

19 Scientific Research Based Optimisation and Geo-information Technologies 361

insists on the importance and the necessity of an intelligent system based on thescientific research and technology development for disaster management and riskreduction.

To achieve an efficient solution to disaster risk reduction, this chapter links widerstrategic aims of bringing together innovative methods based on many scientificdisciplines (e.g., geo-information technology, earth observation techniques, artificialintelligence, early waning systems, dynamic optimisation, risk analysis and environ-mental impact assessment, spatial and environmental planning, etc). More precisely,the purpose of this chapter is to implement robotic algorithms for providing auto-mated data processing strategies to find optimal solutions to disaster managementand risk reduction. This will provide access to a wide range of data collected at aninvestigated region, and combine the observational data with practical data analy-sis in order to improve forecasting and risk assessment. This chapter highlights thecritical role of technology in disaster reduction and management and identifies afew key areas for strengthening/improving technological inputs to the operationalsystem. Section 2 discusses the disaster management cycle and presents all thepractical activities that must be carried out during all the phases of this cycle to min-imise the disaster risk reduction. Also, it outlines disaster management proceduresand components of the hazards analysis for risk assessment. Section 3 presents themost recent processes that have been made through advances in early warning andobserving systems, computing and communications, scientific research and discov-eries in earth science, and how this is helping to understand the physics of hazardsand promote integrated observation and modelling of the disaster. Furthermore, itdiscusses the use of scientific research and technology development in supportingdecision support systems using early warnings for disaster risk reduction. Section 4outlines the disaster warning network and its real-life applications which utilises thestrengths of geo-information technology, dynamic optimisation, information com-munication technology, the internet for providing and representing spatial data, anddynamic models for analysing temporal processes that control the disaster. Section 5describes the geo-information technologies that support and accelerate the searchprocess during all the phases of the disaster. Also, it presents the role of these tech-nologies and other advanced methods during the operational process for creatingdigital maps for disaster management. Section 6 shows the important part of infor-mation communication technology and other supporting tools in accelerating theinformation flow during the phases of the disaster management cycle. Section 7outlines the framework for developing a dynamic model of the disaster monitoringnetwork, and it describes the structure of the central database that will be connectedto this network. Also, it explains optimisation metaheuristic techniques that will beincluded in the dynamic model to accelerate the search process for achieving earlywarring that support the decision support system. Section 8 illustrates some real-lifeapplications based on the use of the disaster warning network, and it insists on theimportance of the capacity building in achieving successful use of all the abovetechnologies for risk reduction. This chapter ends with some recommendations,conclusions and future work.


2 Disaster Management

Disasters are tragic events to development process as they end lives, disrupt socialnetworks, and destroy economic activities. They cut across many organizational,political, geographic, professional, topical and sociological boundaries (Turner andPidgeon, 1997). Therefore, there is a need to integrate information and knowledgeacross many disciplines, organizations, and geographical regions. An effective andcomprehensive disaster management system must allow access to many differentkinds of information at multiple levels at many points of time. It is a continuousprocess by which all individuals, groups, and communities manage hazards in aneffort to avoid or minimize the impact of disasters. Several exact interconnectingsteps (depending on the disaster phase) are typically required to generate the type ofaction that is needed by the disaster management community. Disaster managementinvolves preparing, warning, supporting, and then rebuilding society when disastersoccur. More specifically, it requires response, incident mapping, establishment ofpriorities, and the development and implementation of action plans to protect lives,property, and environment (Cuny, 1983). The following sections present the generalframework for the disaster management cycle and the phases that differ accordingto the type of the disaster.

2.1 The Disaster Management Phases

Disaster management activities, generally, can be grouped into six main phases thatare related by time and functions to all types of emergencies and disasters. Thesephases are also related to each other, and each involves different types of skills anddata from a variety of sources. The appropriate data has to be gathered, organized,and displayed logically to determine the size and scope of disaster managementprograms. During an actual disaster, it is critical to have the right data displayedlogically, at the right time, to respond and take appropriate action for emergencies(Mileti, 1999). Figure 19.1 depicts the framework for the disaster management cyclewhich consists of six phases:

The Prevention and Mitigation phase includes the activities that are trying toprevent a disaster and minimize the possibility of its occurrence (e.g., legislationthat requires building codes in earthquake prone areas, implementing legislationthat limits building in earthquake or flood zones, target fire-safe roofing materials inwild land fire hazard areas, etc). These actions are designed to reduce the long-termeffects of unavoidable disasters (e.g., land use management, building restrictions inpotential flood zones, etc). When potential disaster situations are identified, miti-gation actions can be determined and prioritized. For example, in the case of anearthquake, some questions must be examined: What developments are within theprimary impact zone of earthquake faults? What facilities are in high hazard areas(main bridges, primary roads, hospitals, hazardous material storage facilities, etc.)?What facilities require reinforced construction or relocation? Based on the expectedmagnitude of an earthquake, the characteristics of soils, and other geologic data,what damage may occur? (Handmer and Choong, 2006).


5) Rehabilitation & Reconstruction - Recovery and Early damage assessment - Re-establishing life-lines transport & communication infrastructure - Reinforcement

6) Post Disaster - Lessons learned - Scenario update - Socio-economic & environment impact assessment - Spatial (re)planning

1) Prevention & Mitigation - Hazard prediction & modelling. - Risk assessment & mapping - Spatial planning - Structural & non-structural measures - Public Awareness & Education

4) Response -Dispatching of resources - Emergency telecom - Situational awareness - Command control coordination - Information dissemination - Emergency healthcare

3) Alert -Real time monitoring & forecasting -Early warning -Secure & dependable telecom -Scenario identification -All media alarm

2) Preparedness -Scenarios development - Emergency Planning - Training - Food & medical supplies

Disaster

Information Communication

Technology ICT

is used during most of all the phases of the

disaster cycle

Fig. 19.1 The disaster management cycle

The Preparedness phase includes plans or preparations made and developed bygovernments, organizations, and individuals to save lives, property, and minimizedisaster damage (e.g., mounting training exercises, installing early warning systems,and preparing predetermined emergency response forces). In addition, these activ-ities seek to enhance and help the disaster response and rescue service operations(e.g., providing vital food and medical supplies, mobilizing emergency responsepersonnel, and training exercises).

The Alert phase supports all early warning processes such as real time monitor-ing and forecasting, secure and dependable telecom, scenario identification, and allmedia alarms.

The Response phase is the implementation of action plans that including activ-ities for following the disaster, to provide emergency assistance for causalities andsave lives (e.g., search and rescue, emergency shelter, and medical care, and massfeeding) to prevent property damage, preserving the environment (e.g., shuttingoff contaminated water supply sources), and speeding recovery operations (e.g.,damage assessment).

The Rehabilitation and Reconstruction phase starts when the disaster is over andincludes activities that assist a community to recover and return to normality aftera disaster was occurred. These activities are divided into main two sets: short-termrecovery activities that restore vital services and return vital life support systems tominimum operating standards (e.g., clean up, ensuring injured people have medical


care, providing temporary housing or shelter to citizens who have lost homes inthe disaster, and access to food and water), and long-term recovery activities thatmay continue for several years after a disaster and return life to normal or evenimproved levels (e.g., community planning, replacement of homes, water systems,streets, hospitals, bridges, schools, etc.) (Wisner et al., 2004).

The Post Disaster phase includes analyzing lessons learned, scenario updates,socio-economic and environment impact assessment, and spatial re-planning, etc.These six phases usually overlap, as the information communication technology isbeing used in all these phases, but the usage is more apparent in some phases thanin the others (Ramesh et al., 2007).

2.2 Disaster Management Planning and Hazards Analysis

Disaster management programs are developed and implemented through the anal-ysis of information, most of which is spatial, and therefore can be mapped. Oncethis information is mapped and data is linked to the map, disaster managementplanning activities can begin. These activities are necessary to analyze and doc-ument the possibility of a disaster and the potential consequences or impacts onlife, property, and the environment. This includes assessing hazards, risks, mitiga-tion, preparedness, response, and recovery needs. Planning disaster managementstarts with locating and identifying potential disasters using advanced technology.For example using a GIS, officials can pinpoint hazards (e.g., earthquake faults, firehazard areas, flood zones, shoreline exposure, etc.) and begin to evaluate the conse-quences of potential emergencies or disasters. When hazards are viewed with othermap data (e.g., streets, pipelines, buildings, residential areas, power lines, storagefacilities, etc.), emergency management officials can begin to determine mitigation,preparedness, response, and possible recovery needs. Public safety personnel canfocus on determining where mitigation efforts will be necessary, where prepared-ness efforts must be focused, and where response efforts must be strengthened, andthe type of recovery efforts that may be necessary. Before an effective emergencymanagement program can be implemented, thorough analysis and planning must bedone. GIS facilitates this planning process by allowing planners to view the appro-priate combinations of spatial data through computer-generated maps. This will beexplained in more detail in Sect. 5. Once life, property, and environmental values arecombined with hazards, emergency management personnel can begin to formulateall the disaster cycle: prevention and mitigation, preparedness, alert, response, reha-bilitation and reconstruction, and post disaster plans needs. Important componentsof these plans are mapping hazardous areas, analyzing potential risks to the com-munities and the individuals, and estimating possible losses and damages resultingfrom the disasters.

The quality of spatial and attribute data plays an important role in achievinga successful hazard analysis which aims to identify properties and populationswithin a region that are most at risk from natural disasters. The hazard anal-ysis usually includes five components: hazard identification, profiles of hazard


events, community profile, estimating losses, and vulnerability analysis. The hazardidentification is to identify which types of natural disasters that have the potentialof occurring within a region, and in this case, recorded incidences of past naturaldisasters were used to make this determination. Profiles of hazards events identifypast incidences of natural disasters within each region. In this part, the informationand data presented in these profiles were obtained through review of historical datafrom news media sources, and discussions with community residents and officials.The community profile then compares overall county property statistics to thosewithin the pertinent hazard area. In the last stage of the hazard analysis, individualparcels and property asset data were used in the determination of estimated lossesand vulnerability analysis. Also, advanced geo-information technology (especiallyGIS) is an ideal tool to fulfill all the above tasks of hazard analysis as shownin Sect. 5.

3 Real-Time Early Disaster Warning Network

Space technologies provide valuable tools for the solution of many real-worldproblems in fields such as weather forecasting, communication, and disaster man-agement. With satellite communications, people sending or receiving informationdo not have to be connected to a ground network. With ground-based networks,satellite communications provide access to much of the information over the WorldWide Web (Internet). However, there are weak points in operational utilisationof these technologies, such as inadequate coverage of space data, the effects ofclouds on optical data, inadequate terrain models, assimilation of data in modelsetc. An ideal system needs to have sub-systems on vulnerability/risk assessment,early warning and monitoring, emergency communication and short/long term mit-igation strategies. Therefore, in recent years, the focus of disaster managementcommunity is increasingly moving to the more effective utilization of these tech-nologies, enabling communities at risk to prepare for, and to mitigate the potentialdamages likely to be caused due to the natural disasters. Using these advanced andhybrid technologies, the main application to be considered as a warning base forall the disasters is the designing of a geomatic network which implements a setof control stations spread over the whole geographic area of the hazardous region.The network provides reliable information on a continuous basis through the par-allel process of coverage accuracy prediction (using Least-Squares equations) andintegrity risk simulation functions (using Monte Carlo sampling). The major part ofthe above processes was successfully demonstrated in simulation software consid-ering all standard ranging errors (e.g., satellite clock, ephemeris, multipath, receivernoise, troposphere and ionosphere, etc.) (Saleh, 1996). Then, this network was inte-grated with a dynamic model based on advanced metaheuristic algorithms (whichare based on ideas of Artificial Intelligence) to find the optimal design for this net-work as shown in the designing geomatic networks of Malta and Seychelles (Salehand Dare, 2001, 2002b).


Geo-Information Technology (GIS, GPS, RS, processing etc)

Central Database

Data processing and analytical

centre

User interface (Optimisation,Forecasting, Simulation, Modelling)

Fig. 19.2 The real-time warning network and its database structure

The developed warning network utilizes the strengths of the most advanced geo-information technologies such as geographic information systems and centralizeddatabases, remote sensing, global navigation satellite systems, dynamic optimisa-tion and geospatial models, data collection, hand-held GPS, internet, networking,information communication technology and service delivery mechanisms, warningdissemination, expert analysis systems, information resources etc. This will havepotential to provide valuable support to decision making through providing andrepresenting spatial data, and dynamic models in analysing and representing tem-poral processes that control the disasters. The combined system of the network anddynamic model will be connected to the central database that combine environ-mental and geophysical data from earth observation, satellite positioning systems,in-situ sensors and geo-referenced information with advanced computer simula-tion and graphical visualisation methods as shown Fig. 19.2. Hence, the databasewill provide the following internet-based services: quickly locating and ensure dataavailability where and when needed, detailed descriptions of the contents and limita-tions of the data, and presenting the data in different formats (maps, graphs, pictures,videos, etc.). In addition, the database will be designed to be searchable (by datatype, data holder/owner, location, etc), and will be used in three modes: planningand design for protection, real-time emergency, and disaster recovery (Saleh, 2003).

4 Scientific Research and Technology Developmentfor Early Warning

Exchange of information and communication practices play key roles in the real-ization of effective disaster management and risk reduction activities. Therefore,operational use of technology, in terms of information gathering and their real


time dissemination leading to effective risk reduction at the national and locallevel, requires appropriate facilities, techniques and institutional systems to bein place. In general form, the disaster information system includes three sub-systems: knowledge, interconnectivity, and integration. Knowledge sub-systeminvolves observation techniques for data collection and visualization, analysis, fore-casting, modelling and information management. The interconnectivity sub-systemrelates to the mode of communication employed to retrieve and distribute data and tothe dissemination of information products. On the other hand, the integration sub-system addresses the operational system, standards and protocols, procedures forevaluation of quality and reliability and training of key personal. Data availabilityis crucial for ongoing research, to monitor hazards and to assess risks. Integratingnew developments in information management with established and more tradi-tional methods can help to create a better understanding about hazards and theirrisks. Effective information management and communication are also instrumentalfor Early Warning (EW) systems and effective mitigation efforts. The main objec-tives of EW is to be better prepared to face challenges of the risk of short/longterm or sudden disasters through these steps: (1) avoiding and reducing damagesand loss, (2) saving human lives, health, economic development and cultural her-itage, and (3) upgrading quality of life. However, the main EW challenges can beseen when: (1) risks and warnings are not understood, (2) information is scattered,and (3) dissemination is limited. Within this context of EW, the main purposes ofscientific research is to overcome these challenges which can be summarised andconcluded as so: (1) bridge gaps between science and decision making communities,(2) increased warning time/quicker response, and (3) better understand disasters.Therefore, scientific research, geo-information technology, forecasting, modelling,warning systems are only valuable when they are applied and when they are put intopractice for disaster management as shown in Fig. 19.3. Taking this in considera-tion, the optimal decision support system can be achieved through the following:(1) integrating information, science, research, and technology to improve decisionsupport capabilities, (2) improved observation systems/data/analysis, (3) advancedalgorithms, and models, (4) GNSSs, RS, GIS, visualization and display, (5) ICT andnetworks (EEA, 2001).

More practically, an effective early warning system (which must be concen-trated on the people at risk) is consisted of four main parts: risk knowledge andassessment; technical monitoring and warning service; dissemination of warnings;and public awareness and preparedness (Egeland, 2006). These main parts must beintegrated in one system and failure in any one of them will cause failure of thewhole early warning system. To achieve this effectiveness for this early system,significant progress and large improvements have been made in the quality, time-liness and lead time of hazard warnings and decision support system for disastermanagement and risk reduction. All these advances have been marked and drivenby scientific research and technology development particularly through the use ofthe computer sciences, artificial intelligence, and operational research, etc. on theother hand, there have been continuous improvements in the accuracy and reliabilityof monitoring instrumentation, and in integrated observation networks particularly


Modelling

Operational Research

Simulation

Optimization

Best Practices

RemoteSensing

GNSSs

GISs

Internet & Intranet

ICT

All types of

Surveying Methods

EnvironmentalDisasters

Natural Disasters

Man-madeDisasters

Applications of new technologies in

supporting Early Warning System & Decision Support

System for disaster management &

risk reduction

Disaster Management & Risk Reduction

Geo - information & Communication

Technology

Scientific Research & Technology

Development

Artificial Intelligence

Fig. 19.3 The complete system of scientific research and technology development for disastermanagement and risk reduction

through the use of geo-information and information communication technologies,internet, and other observation methods. As shown in Fig. 19.3, these developments,in turn, have supported research on hazard phenomena, modelling, simulation, mon-itoring, detecting and forecasting methods and developing hazard warnings for awide range of all types of natural and man-made disasters.

However, capacities in the monitoring and prediction of these disasters vary con-siderably from one disaster to another and are faced by major challenges and gaps.Some of these challenges and gaps include the availability of these technical capa-bilities and its integration into the disaster risk reduction decision process within asustainable procedure; the need for improvement of technical warning capabilitiesfor many hazards. Other gaps and challenges can be seen in insufficient coverageand sustainability of observing systems for monitoring of all type of disasters andhazards, insufficient level of technical capabilities (resources, expertise and opera-tional warning services) in the operational technical agencies that are responsible for


monitoring and forecasting of severe hazards, difficult access to information fromthe related teams outside of the affected areas, weak communications for providingtimely, accurate and meaningful forecasting and early warning information to all theusers.

With regards to dissemination, telecommunication mechanisms must be opera-tional, robust, and available every minute of every day, and tailored to the needs ofa wide range of threats and users. All of these mechanisms must be based on clearprotocols and procedures and supported by an adequate communications infrastruc-ture. However, there are gaps and challenges that affect warning messages due tothe underdeveloped dissemination infrastructure and systems, the incomplete cov-erage of systems, and the resource constraints contributing to the lack of necessaryredundancy in services for information. Other gaps and challenges might includeinsufficient institutional structures to issue warnings due to limited understanding ofthe true nature of early warning, lack of clarity and completeness in warnings issueddue to the lack of common standards for developing warning messages within andacross countries, unclear responsibilities about who provides forecasts (of hazards)and who provides warnings (of risks), insufficient understanding of vulnerabilitydue to the lack of better integration of risk assessment and knowledge in the author-itative, official warnings at the national level, ineffective engagement of warningauthorities with the media and private sector, (6) the lack of feedback on the systemand its performance and learning from previous experience.

The characteristics of risk can usually be presented through scenario plans, prac-tical exercises, risk mapping, and qualitative measures, etc. To improve the basis forcollecting and analysing risk data, risk assessment requires standard indicators tomeasure the success and failure of early warning systems. Therefore, the develop-ment of effective warning messages must depend on relatively good data resourcesand the generation of accurate risk scenarios showing the potential impacts ofhazards on vulnerable parts of hazardous area. In this direction, more researchis needed to make qualitative data and narratives of vulnerability accessible anduseable to engineers, planners, policy makers and all the other parties working inthis domain. This will support the rescue teams with capabilities to analyse notonly the hazards, but also the vulnerabilities to the hazards and the consequentsof the risk, and thus will help them decide whether and when to warn. In addi-tion to gathering statistics and mapping populations’ risk factors, risk assessmentsshould involve the community to ascertain their perceived risks and concerns. Toensure the optimal decision support system for natural and environmental disastermanagement and risk reduction using early warning capabilities, capacity buildinghas to be highly considered on all the aspects as follows: Academic programmeand technical workshops: training of scientist and engineers during installationphase. Institutional capacity building: consulting of organizational structures andinter-institutional communication, planning and construction of new infrastructures,establishment of communication platforms and chains. Warning culture: establish-ment of warning mechanisms products (e.g., risk maps, evacuation plans, etc.) andinformation products for end users, (e.g. development of teaching units in schools,universities, and the community).


5 Geo-Information Technology

Geo-information technologies provide real-time information that allows agenciesworking on disaster management and risk reduction to effectively manage the situa-tion and to plan community evacuation and relief operations in case of emergencies.These technologies can help considerably to show vulnerable areas, enhance map-ping, and ameliorate the understanding of hazards (Oosterom et al., 2005). Thefollowing sub-sections present these advanced technologies and their roles indisaster risk reduction.

5.1 Global Navigation Satellite Systems (GNSSs)

It is well known that throughout the world the use of the GNSSs is dramati-cally increasing, demanding the optimisation of the accuracy of the measurementsprovided by these systems. The Global Positioning System (GPS), the GLObalNAvigation Satellite System (GLONASS), and the European Satellite Navigation(Galileo) are the most widely known satellite systems as shown in Fig. 19.4. GNSSsSatellites provide the user with a 24-h highly accurate three-dimensional posi-tion, velocity and timing system at almost any global location. However, thesesystems suffer from several errors that affect the accuracy of the observationand Fig. 19.5 depicts the sources of these errors. Large part of these errors canbe theoretically and practically minimized and eliminated and using differentialand wide area augmentation systems and other surveying methods (Elliott, 1996)and (Leick, 1995).

5.2 Remote Sensing (RS)

RS satellites are used to monitor the land, the surface, the oceans and the atmo-sphere, and how their situations they change over time. Most RS satellites cover the

Fig. 19.4 The GPS and GLONASS constellation of navigation satellite systems


Selective AvailabilityAtmospheric Delays

Receiver ClocksMultipath Delays

Satellite ClocksEphemeris

Fig. 19.5 Sources of errorsin GNSSs

whole globe, making them important for the study of large-scale phenomena suchas climate changes and desertification as shown in Fig. 19.6. For example, remotelysensed imagery helps to identify the most fire-prone areas and to develop fire prop-agation models which allow emergency evacuation to be modeled at the level of theindividual vehicle for avoiding congestion during evacuation. In addition, RS hasapplication in the characterisation of earthquake risk through the identification ofregions prone to liquefaction (river valleys and coastal areas). Seismic vulnerabilityfrom tsunamis is easily assessed by convolving digital elevations and bathymetrydata with the distribution of coastal populations and economic infrastructures.However, the main limitations of RS satellite images are cloud cover and resolution.

Fig. 19.6 The remotesensing technology


Some of these problems may be circumvented using GNSSs satellites. Combiningremotely sensed imagery with ground data reduces the cost of ground-based sam-pling efforts by more than 50% while substantially increasing the accuracy ofcollected data (Brown and Fingas, 2001).

5.3 Geographic Information Systems (GISs)

Innovations in GIS technology are increasingly accepted tools for the presentation ofhazard vulnerabilities and risks. The data obtained from GNSSs and RS will be usedby GISs technology to produce maps that identify and analyze all applicable typesof natural hazards. These maps then can be used by local governments to informcitizens within their communities of the potential risks from these hazards. GISsfacilitate the integration of data obtained from various sources (e.g., topographichardcopy maps, tables, aerial photos, satellite images, satellite navigation systems,etc). Then, this data will be analysed and processed to produce “Smart Maps” thatlink database to map and for every feature on this map, there is a row in a table.Figure 19.7 depicts the GIS operational cycle to process geographic information andcreate digital maps through these steps: data acquisition, data processing, and datadissemination. By utilizing a GIS, all related parties can share information throughdatabases on computer-generated maps in one location. GISs provide a mechanismto centralize and visually display critical information during a disaster (Masser andMontoya, 2002).

5.3.1 The Role of GIS During the Disaster Management Phases

GIS plays an important part during all the disaster management phases as explainedpreviously in Sect. 2. The Planning phase for disaster involves predicting thearea and time of a possible disaster and the impacts on human life, property, and

Sources of Geographic InformationData Processing

& Modelling Visualising Geographic Information

Database“not easy to interpret”

Visualisation “worth a thousand words”

Fig. 19.7 The operational cycle for geo-information technologies to create digital maps


environment. These factors are used to determine an effective planning procedurefor the mitigation of possible disaster effects. This planning can be done effectivelyand quickly using the application of GIS, which is a very good tool for short andlong term planning. GIS modeling allows disaster managers to view the scope anddimension of a disaster and its impacts. GIS allows disaster managers to quicklyaccess and visually display critical information by location. This information facil-itates the development of action plans that may be transmitted to disaster responsepersonnel for the coordination and implementation of emergency initiatives.

During the mitigation and prevention phase (and by utilizing existing databaseslinked to geographic features), GIS can be used for managing large volumes ofdata needed for the hazard and risk assessment as values at risk can be displayedquickly and efficiently through a GIS. For example, in case of a wildfire disaster,a GIS can identify specific slope categories in combination with certain species offlammable vegetation near homes that could be threatened by wildfire. GISs cananswer these questions: Where are the fire hazard zones? What combination of fea-tures (for example, topography, vegetation, weather) constitutes a fire hazard? Withregards to the other disasters such as earthquakes and floods, a GIS can identifycertain soil types in and adjacent to earthquake impact zones where bridges or over-passes are at risk. GISs can identify the likely path of a flood based on topographicfeatures or the spread of a coastal oil spill based on currents and wind. Most impor-tantly, human life and other values (property, habitat, wildlife, etc.) at risk can bequickly identified and targeted for protective action.

During the preparedness phase, GISs can be use as a tool for planning of evacua-tion routes, for the design of centers for emergency operations, and for integration ofsatellite data with other relevant data in the design of disaster warning system. Theycan provide answers to questions to those activities that prepare for actual emer-gencies: Where should fire stations be located if a short response time is expected?How many paramedic units are required and where should they be located? Whatevacuation routes should be selected? How will people be notified? Will the roadnetworks handle the traffic? What facilities will provide evacuation shelters? Whatquantity of supplies will be required at each shelter?

For Early warning purposes, GISs can display real-time monitoring for earlyemergency warning. Remote weather stations can provide current weather indexesbased on location and surrounding areas. Wind direction, temperature, and relativehumidity can be displayed by the reporting weather station. Wind information isvital in predicting the movement of a chemical cloud release or anticipating thedirection of wildfire spread upon early report. Earth movements (earthquake), reser-voir level at dam sights, and radiation monitors can all be monitored and displayedby location. It is now possible to deliver this type of information and geographicdisplay over the Internet for public information or the Intranet for organizationalinformation delivery.

During the response phase, the closest (quickest) response units based at fixedand known locations can be selected, routed, and dispatched to a disaster. Dependingon the kind of the disaster, GISs can provide detailed information before the firstunits arrive. For example, during a fire in housing area and while the rescue team


in the route to the emergency, it is possible to identify the closest hydrants, electri-cal panels, hazardous materials, and floor plan of the building. For hazardous spillsor chemical cloud release, the direction and speed of movement can be modeledto determine evacuation zones and containment needs. Advanced vehicle locatingcan be built-in to track (in real-time) the location of incoming emergency units andthen to assist in determining the closest mobile units (which are located on the mapthrough GNSS transponders) to be dispatched to a disaster. During multiple dis-asters (numerous wildfires, mud slides, earthquake damage) in different locations,GISs can display the current emergency unit locations and assigned responsibilitiesto maintain overall situation status. In general, the response phase is divided intotwo phases: a short-term phase and a long-term phase. One of the most difficult tasksin the short-term recovery phase is damage assessment, but a GIS integrated withGNSSs can play important roles such locating each damaged facility, identifying thetype and amount of damage, displaying the number of shelters needed and wherethey should be located for reasonable access, and displaying areas where serviceshave been restored in order to quickly reallocate recovery work to priority tasks.In this phase, laptop computers can update the primary database from remote loca-tions through a variety of methods. GISs can display (through the primary database)overall current damage assessment as it is conducted. Emergency distribution cen-ters’ supplies (medical, food, water, clothing, etc.) can be assigned in appropriateamounts to shelters based on the amount and type of damage in each area. Actionplans with maps can be printed, outlining work for each specific area. Shelters canupdate inventory databases allowing the primary command center to consolidatesupply orders for all shelters. The immediate recovery efforts can be visually dis-played and quickly updated until short term recovery is complete. This visual statusmap can be accessed and viewed from remote locations. This is particularly help-ful for large emergencies or disasters where work is ongoing in different locations.During the long-term recovery phase, prioritization plans and progress for majorrestoration investments can be made and tracked utilizing GIS. In addition, responserequirements, protection needs (e.g., supportive bridge in the event of floods, remov-ing vegetation in the case of wildfire, etc) can be determined for areas at high risk.Long term recovery costs can be highly expensive for large disasters and accountingfor how and where funds are allocated is demanding. In this part and after allocat-ing the funds for repairs, accounting information can be recorded and linked toeach location, then GISs can be implemented to ease the burden of accounting forthese costs. In the disaster relief phase, GISs are extremely useful in combinationwith GNSSs in search and rescue operations in areas that have been devastated andwhere it is difficult to orientate. In disaster rehabilitation phase, GISs are used toorganise the damage information and the post disaster census information and in theevaluation of sites for reconstruction.

5.3.2 HAZUS (Hazards U.S.)

While GISs are used to capture, analyse, and display spatial data, the models providethe tools for complex and dynamic analysis. Input for spatially distributes models,as well as their output, can be treated as map overlays (Fedra, 1994). The familiar


format of maps supports the understanding of model results, but provides also aconvenient interface to spatially referenced data. Expert systems, simulation andoptimisation models add the possibility for complex, and dynamic analysis to theGIS. Recently, a new program based on GIS was presented called HAZUS (HazardsU.S.), which is open source, free to use, and highly responsive to end-user require-ments. Users incorporate data and modelling the physical world of infrastructure,build inventory, geology, damage estimation formulas, and critical operating centrelocations, and then subject HAZUS to the complex consequences of a hazard eventas shown in Fig. 19.8. After that, users can implement HAZUS to prepare for dis-asters (pre-event), respond to the threat (during the event), analyze and estimate thepotential loss of life, injuries, property damage, forecasts casualties, and to managethe critical situation (post-event).

One major challenge in building effective information systems for disaster man-agement (e.g., fault movement, river basin) is the integration of dynamic modelswith the capabilities of GIS technology. This can provide a common framework ofreference for various tools and models addressing a range of problems in river basinmanagement, supply distributed data to the models, and assist in the visualisation ofspatial model results in the form of topical maps (Fedra, 1995). The possibility ofapplying HAZUS program to investigate some critical situations in Syria and neigh-bouring countries (e.g., the West Shaam fault as shown in Fig. 19.9) were plannedfor fault extends for about 1,100 km along the western part of the Shamm countries(Syria, Lebanon, Jordan, Palestine) representing the north-western Arabian Africanplates boundary.

Fig. 19.8 HAZUS in estimating the peak ground acceleration and source in the earthquakescenario


Fig. 19.9 The West Shaamfault

6 Information Communication Technology (ICT)for Disaster Management

ICT is used in almost all phases of the disaster management process and can effec-tively be used to minimize the impacts of disasters in many ways. ICT plays acritical role in facilitating the reconstruction process and in coordinating the returnof those displaced by disasters to their original homes and communities. Disastermanagement activities, in the immediate aftermath of a disaster, can be made moreeffective by the use of appropriate ICT tools. These include tools for resourcemanagement and tracking, communication under emergency situations (e.g. use ofInternet communications), and collecting essential items for the victims. GISs andRS are examples of ICT tools being widely used in almost all the phases of disastermanagement activities. RS for early warning is made possible by various availabletechnologies, including telecommunication satellites, radar, telemetry and meteo-rology. More clearly, the rule of used ICT is to accelerate the flow of informationduring all the stages of the disaster between the emergency and rescue teams in dis-aster location and the main authorities (decision makers) in central control room.Any one or a combination of the following ICT and media tools that are shownin Fig. 19.10, can be used in disaster management: radio and television, telephone


Radio Tower

Satellite dish

Telephone

FaxFax Modem

Modem Television

Bulletin

GIS data

Radiowireless communication

via modemmanual entry

Telemetry/Data box/ Voice

Satellite images

Weather forecast

Synoptic charts

Boundary estimationRainfall,Water level

Data Entry&

Processing

Modelling&

Mapping

River stage

Rainfall

Data

Dissemination to the public

Dissemination to various agencies

Internet

GPRSStation

TCP/IPfor

GPRS

GSM Digital Camerawith GPS

Fig. 19.10 The ICT system used for flood monitoring network

(fixed and mobile), short message service (SMS), cell broadcasting, satellite radio,internet/email, amateur radio and community radio (Wattegama, 2007).

7 The Dynamic Metaheuristic Model

Within the concept of dynamic optimisation, these disasters can be regarded asnon-differentiable and real-time Multi-objective Optimisation Problems (MOPs).These problems involve multiple, conflicting objectives in a highly complex searchdomain. Moreover, the volume of data collected for these problems is growingrapidly and sophisticated means to optimise this volume in a consistent and eco-nomical procedure are essential. Therefore, robotic algorithms are required to dealsimultaneously with several types of processes which are concerned with the unpre-dictable environment of these problems (Deb, 2001). These algorithms can providea degree of functionality for spatial representation and flexibility suitable for quicklycreating real-time optimal solutions that account for the uncertainty present in thechanging environment of these problems which can be formulated in a design modelfor the monitoring network as follow in Eq. (1):

NetworkMOP = optimize : f (x) = { f1(x), f2(x), . . . , f2(x)} subject to

x = (x1, x2, . . . , xn) ∈ X(1)


where fi(x) is the model of the network that consists of ith monitoring objectivefunctions to be optimised, x is a set of variables (i.e., decision parameters) and X isthe search domain. The term “optimise” means finding the ideal network in whicheach objective function corresponds to the best possible value by considering thepartial fulfilment of each of the objects. More specifically, this network is optimalin a way such that no other networks in the search domain are superior to it whenall objectives are considered.

The main innovative aspect of the developed network is the integration of thestate of the art geographical and environmental data collection, and data manage-ment tools with simulation and decision tools for disaster management and riskreduction. Then, this network was integrated with the artificial intelligence optimi-sation algorithms to find the optimal network design. This will allow the modeller todevelop a precise and unambiguous specification that can strongly help in estimat-ing the impacts of an actual development process of the presented design. Therefore,it is almost impossible even for an experienced and higher-level designer to findan optimal design by the currently used methods which do not provide spatialrepresentation to the whole situation and lack the ability to select “interesting” con-tingencies for which to optimise. Once such designs are obtained, the technical userwill be able to select an acceptable design by trading off the competing objectivesagainst each other and with further considerations. The final design of the networkshould be robust (i.e., performs well over a wide range of environment conditions),sustainable (i.e., not only optimal under current condition, but also considering pre-dicted changes), and flexible (i.e., allows easy adaptation after the environment haschanged) (Peng et al., 2002).

Initial Network Formation

Neighbourhood (set of alternativenetworks derived from the initial one)

Search by Move Formation (Local Search)

Provisional Neighbourhood Formation

Search by Network Formation(development and guided search techniques)

New Neighbourhood (solution formation)

Acceptance Criteria

Termination of the Search

Fig. 19.11 The generalframework for metaheuristicalgorithms


Metaheuristic techniques (which are based on the ideas of artificial intelligence)potentially have these capabilities to produce set of high quality real-time designsthat can model more closely and easily many functions and visualize the trade-offsbetween them and then to filter and cluster top optimal solution (Osman and Kelly,1992). These techniques are iterative procedures that combine different opera-tional and organizational strategies based on robustness and computerized models inorder to obtain high-quality solutions to complex optimization problems. They canprovide instantaneous comparisons of the achieved results of different developeddesigns using several procedures such as convergence, diversity, and complexityanalysis, etc. Figure 19.11 depicts the general framework of the metaheuristics algo-rithms that has been adopted in this research. The dotted lines indicted option thatcan be skipped or used. In this research, several metaheuristics are proposed andimplemented for optimising the scheduling activities of designing the monitoringnetwork. The well-known metaheuristics that have been successfully applied to opti-mise real-life applications based on monitoring network are: simulated annealing,tabu search, ant colony optimization, and genetic algorithm (Saleh and Dare, 2002a).These metaheuristics are inspired, respectively, by the physical annealing process,the proper use of memory structures, the observation of real ant colonies and theDarwinian evolutionary process.

7.1 Simulated Annealing (SA)

The SA technique is flexible, robust and capable of producing the best solution tocomplex real life problems (Kirkpatrick et al., 1983) and (Rene Vidal, 1993). Thistechnique derives from physical science and is based on a randomisation mechanismin creating solutions and accepting the best one. The annealing parameters thathave to be specified are; the initial temperature, the temperature update function,the length of the Markov chain and the stopping criterion. The initial temperaturesimulates the effect of temperature in the search process to find the best candidateof the final design. The temperature update function determines the behaviour of thecooling process, while the length of the Markov chain represents the number of iter-ations between the successive decreases of temperature. The optimization process isterminated at a temperature low enough to ensure that no further improvement canbe expected. With a suitable annealing parameters, an optimal network design orclose to it can be achieved for optimization the flooding problem (Saleh and Allaert,2008). The basic steps for the SA, which returns a better network, are depicted inFig. 19.12.

7.2 Tabu Search (TS)

The TS technique, which is a global iterative optimisation, exploits knowledge ofthe system or “memory” under investigation to find better ways to save compu-tational efforts without effecting solution quality (Glover and Laguna, 1997). The


Initialisation: Step 1 Select an initial Network design candidate NINT with value V(NINT) Step 2 Initialise the annealing parameters:

• Set the initial temperature Ti. • Set the Markov temperature length L. • Set the cooling factor F (F<1). • Set the iteration counter K. Selection and acceptance strategy of the generated neighbouring candidates: Step 3 Generate I(NINT) neighbours {N1,..,Nn} of the NINT and compute the value V(NINT).

Step 4 Select the best possible candidate with value V(N’) , N’∈ I(NINT). • Compute Δ= V(N’)–V(NINT)• IF {Δ≤ 0 or e−Δ/ T >θ} where θ is a uniform random number 0<θ<1. THEN accept the new candidate N’ as a current one NCN and set NCN → NINT OTHERWISE retain the NCN. and update the annealing parameters: Step 5 Update the temperature according to rule (Tk+1=F*Tk ) • Set K→K+1 Step 6 If the stopping criterion is met then stop and report the Best Found Candidate NBFN and K. OTHERWISE go to Step 3.

Fig. 19.12 Basic steps of the SA procedure

most basic form of the TS is the construction of a tabu list which prevents the searchfrom cycling by forbidding certain candidates and then directing the search towardsthe global optima. At the beginning of the process, this list is often empty but ismade up during the search process by adding candidates that could return the cur-rent candidate to previous local optima. Implementation of TS requires specificationof tabu parameters: the tabu list, the candidate list, the tabu tenure, and the stoppingcriteria. The tabu list is a memory structure that prohibits moves that have recentlybeen interchanged to prevent cycling. The candidate list contains a set of selectedmoves that gives the best-generated neighbouring candidates surrounding the cur-rent candidate. The tabu tenure determines the number of iterations for which acandidate maintains its tabu status and more information about the selection of thetabu parameters for practical applications can be found in (Saleh and Dare, 2003).

7.3 Ant Colony Optimization (ACO)

The ACO is a multi-agent approach to search and reinforce solutions in order to findthe optimal ones for hard optimization problems. This technique is a biological-inspired agents based on the foraging behaviour of real ant colony for distributedproblems-solving (Dorigo and Gambarddella, 1996). The basic idea underlying thismetaheuristic is the use of chemical cues called pheromone (a form of collectivememory). The function of these pheromones is to provide a sophisticated commu-nication system between ants that cooperate in a mathematical space where theyallowed to search and reinforce pathways (solutions) in order to find the optimal one.This metaheuristic include; positive feedback (intensity to quickly discover goodsolutions), distributed computation (to avoid premature convergence), and the use


of a constructive greedy metaheuristic (visibility to help find an acceptable solutionin the early stage of the search process) (Saleh, 2002).

7.4 Genetic Algorithms (GAs)

Unlike the above mentioned techniques, GAs, which are inspired from popula-tion genetics, operate on a finite pool of solutions (usually called chromosomes)(Goldberg, 1989). The chromosomes are fixed strings with binary values at eachposition. The main idea behind GAs is to maintain this pool of selected solutions thatevolves under selective pressure that favours better solutions. To facilitate producingthese better solutions and preventing from trapping in local optima, a set of geneticoperators are used. These operators include cross-over, mutation, and inversion. Incross-over, some cut-points (members of the population) are chosen randomly andthe information between these chromosomes are exchanged. The mutation operatorprevents GAs from trapping in a local optima by selecting a random position andchanging its value. In Inversion, two-cut points are chosen at random and the orderof the bits is reversed (Saleh and Chelouah, 2003).

8 The Real-Life Applications of the Disaster Warning Network

Some practical examples based on the use of the above developed systems of thegeomatic network and advanced metaheuristic algorithms will be presented andexplained.

8.1 The Handicapped Person Transportation (HPT)Problem in the City of Brussels

An effective method based on the genetic algorithms has been implemented to solvethe handicapped person transportation problem which is a real-life application thatrepresents large part of the disaster management and emergency response activi-ties that can be carried out during and after the disaster to rescue the injured. In thisapplication, vehicles (e.g., ambulances in case of a disaster) have to transport patient(e.g., casualties) from their locations to different destinations (hospitals or recoverycentres). The objective is to maximize the service quality of the transportation byfinding optimal routes for transporting handicapped people while minimizing thenumber of the used vehicles under the constraints of desired times of transportationand vehicle’s capacity. This usually takes the form of constraints or objective func-tion terms related to waiting times and ride times (the time spent by a user in thevehicle) as well as deviations from desired departure and arrival times. The obtainedresults show that the developed HPT software (as shown in Fig. 19.13) can effec-tively provide high-quality solutions in terms of service quality and computationaleffort. This speedup in obtaining good solutions has a significant impact on the


Fig. 19.13 The HPT software

quality of the services that offered to the patients (e.g., casualties). In addition, abetter reactivity of the scheduling operation allows the transportation enterprise tominimize the delay time between travel requests and at the same time to providea better use of the enterprise’s resources. This is successfully applied on real-life instances during working days in the city of Brussels by the Inter-CommunalTransport Company of Brussels (Rekiek et al., 2006).

The HPT software was specifically designed to optimize operations inon_demand bus/minibus transportation enterprises. This software features addi-tional facilities for handling special cases of the HPT problem. Given a set oftransport demands, the software automatically produces routes for buses in the enter-prise’s fleet. Each route gives the chronological list of stops where the bus will carryout together the list of clients entering and leaving the bus at each stop. The soft-ware is currently used by the Belgian public authority for daily scheduling of theHPT problem and can be characterized by the following:

• Origin and destination locations (called stops) where the client must be pickedup and delivered;

• For each trip, the pickup or the delivery of the client must occur at a given desiredtime instant according to his/her request;


• The total time needed to transport a given number of clients must not exceedthe prescribed maximum travel time which is proportional to the time needed totravel directly from origin to destination.

• Providing additional service times at origin and destination (e.g., the time neededto get out the client from the vehicle and to carry him/her into the hospital);

• The service requirements of each client depend on his/her physical characteris-tics.

• A heterogeneous fleet of vehicles is available for the services which made up ofa limited number of minibuses;

• Each vehicle has its own configuration (capacity constraints) and can performmore than one transportation request during a day;

• At any time instant, the number of served clients must not exceed the vehiclecapacity;

• A vehicle is normally a depot-based, therefore it starts its route from the depotand it must return to the same depot at the end of the trip;

• Each request have to be assigned to exactly one vehicle;• A vehicle enters or leaves a location only if an origin or destination of a request

is assigned to that vehicle;

There are two types of trips according to the client that must be serviced at thedesired time instant as follows. In outward trips in which the arrival at the destina-tion stop must occur at the instant dti (e.g., a trip from home to hospital). On returntrips, the departure from the origin stop must occur at the instant oti (e.g., a tripfrom hospital back to home). The HPT problem in this research is time windowsconstraint. Therefore, if the vehicle arrives at the origin stop of a return trip tooearly, it must wait until time instant of the services of the client begins. The vehiclecan only leave once for the next stop to carry the client. Analogously, if the vehiclearrives at the destination stop of an outward trip too early, it must wait until theinstant of given trip begins at the origin location.

8.1.1 Input Data

The HPT software needs the following input data:

Clients: Each client has his/her origin/destination address, departure/return timeinterval, the handicap conditions, and the reason for the travel. Some clientsrequire wheelchairs at all times, others must remain seated in wheelchairsduring movement, and some clients need accompanying personnel. Otherconsideration are:

• The client can not be served before time “no-sooner than” or “no-laterthan” time interval.

• The place is specified by simple street address of the stop and validatedby an automatic connection with the map.

• A detailed client file is maintained by HPT which includes severaladdresses of the client allowing for a speedy input of a stop.


• For the handicapped persons, the client file maintains various data con-cerning the handicap condition (e.g., special facilities are required fortransportation, the typical time necessary for the person to enter andleave the bus, etc.).

Vehicles: A vehicle has multiple time windows which can define all the peri-ods in which this vehicle is available. The representation of the fleet of thetransportation enterprise is accessible through a user-friendly interface. Inaddition, the data include the type of each vehicle and the availability of thevehicle in terms of not-sooner-than/not-later-than time interval. The mainconstraints usually are: vehicles are not available all the day and subjectedto the service plans, drivers have to be changed during the day times, char-acteristics of the available vehicles are: each vehicle’s its depot address, anumber of configurations (maximum number of wheelchairs and number ofnon wheelchair clients), and its departure/return time.

Map: In order to design the vehicle routes, HTP software needs an estimateof travel time between the stops. Instead of using a detailed and expensiveroute map, the software obtains this information from the addresses of thestops using a special model (Jaszkiewicz and Kominek, 2003). This modeltakes into account the day time and the direction of the travel in order toavoid rush hours and massive influx into the city during the morning and theevening periods. Other local exceptions may be specified during the process(e.g. for road works). The map is currently available for the Brussels regionand distances are expressed in meter (m) and times in minutes (min). Also,a user-friendly interface is available for elaboration of the map for other ter-ritories. For each pair (i, j) of locations, the estimated travel time (tij) andthe corresponding distance (dij) are given and measured on the road network.The model uses five kinds of trajectories and takes into account seven inter-vals of time corresponding to slack periods and rush hours as illustrated inFig. 19.14 as follows:

• The two locations are inside a region of Massive Influx (MI),• The first location is situated in the region of the MI, while the second

one in a normal region,• The first location is situated in a normal region, while the second one

in the MI,• The two locations are outside the MI region, but the trajectory crosses

the region of the MI,• The two locations are outside of the MI region, but the trajectory does

not cross the region of MI.

8.1.2 Output Data

The main outputs of HTP software are:

• For each bus, there is a route sheet which contains a detailed list of stops in achronological order.


0

2

3 1

4

LocationDirection of vehicleContour of networkRegion of massif influx

Fig. 19.14 Network flow

• For each stop, there is a sheet which specifies its address, the time allocated forthe bus at this stop, and the identity of the clients entering or leaving the bus.

• The times of departure and arrival to the depot.• Important statistics which are useful for many practical purposes (e.g., number

of times when a client has made the last-minute cancellation of a request, etc).

The obtained results show that the HPT software can effectively provide in a rea-sonable computation time (30 min) high-quality solutions which previously took afull day of three persons time to prepare. This speedup in obtaining good solutionshas a significant impact on the quality of the services that offered to the clients.In this case, the personnel can spend their spare time to improve the contact withthe clients if necessary. In addition, a better reactivity of the scheduling operationallows the transportation enterprise to minimize the delay between travel requestsand at the same time to provide a better use of the enterprise’s resources. Origin anddestination locations of the clients are distributed over 250 physical zones coveringthe whole territory of the city of Brussels. The problem consists of a trip whose ser-vice requirements are 164 clients and 18 vehicles and all the numerical input dataare known in advance.

8.2 The Real-Time Monitoring Network for WaterManagement in Flanders (Belgium)

Another real-life application has been developed to design a monitoring network forwater pollution control and water management in Flanders, Belgium (Saleh et al.,2008). The network and its model will be composed with extra elements and func-tions according to the kind of the problem to be optimised. The design model forthis network requires specific objectives for an efficient and effective monitoringsystem that will address many operational and management requirements related towater quality and quantity parameters.


The developed methodology considers all relevant aspects of flood risk: pre-ventive measures, monitoring and forecasting overflows, water management, earlywarning, simulation and optimisation procedures, etc. This will generate knowledgecontributing to the risk and damage assessment prevention of floods and the designof effective response actions maximising the safety measures. Figure 19.15 depictsthese objectives to optimise monitoring violations of the water quality standards,determining water quality status that help understanding the long and short termtrends of temporal variations, identifying the causes and sources affecting waterquality changes, the use of water quality modelling that support scientific waterquality management, etc. The design network and its value model can be com-posed with extra other monitoring parameters according to the purpose of eachmonitoring network. These additional parameters can be used to: support earlywarning of adverse impact for intended water uses in case of accidental pollu-tion, verify of the effectiveness of pollution control strategies, etc. To meet theabove mentioned objectives, water quality data should be collected at the monitoringstations representing each watershed unit in an investigated region.

Figure 19.16 depicts the real-time monitoring network which is a system ofsatellites and ground stations for providing real-time monitoring to detect impact ofthe pollution sources implicated in water quality changes. This network implementsa set of monitoring stations spread over the whole geographic area of the regionto provide reliable information on a continuous basis (Saleh and Allaert, 2007).This network has been designed to consider several levels of planning and decisionmaking through the management system of spatially referenced data with advancedcomputer simulation, graphical visualisation, and dynamic metaheuristc methods.

Data collection and transmission

Measurement calculation

Warning dissemination

Monitoring & Emergency actions

Flood starts

time

Static Observations

Minimise

maximise

Fig. 19.15 The dynamic model for the flood monitoring network


RiskAssessment

EarlyWarning

Monitoring

DamageAssessment

Distribution to the Users

Provision of Forecasts

InternetIP/http,

HydrologicalStations

Data collection

Dynamic DataProcessor

Desktop GIS Tools

Web site, bulletin, e-mail, fax, radio,telephone, etc.

Main Services (File, Spatial DataEngine, Internet Map, Web, etc)

Fig. 19.16 The real time monitoring network for Flanders region

Several important elements must be integrated into a new strategy that will bepractically embedded in the current model. This strategy is involving aspects of:(1) spatial and environmental planning and land-use regulation (e.g., declarationof flood risk areas as priority and reserve areas, etc), (2) water management (e.g.,determination of flood areas, installation of flood action plans, and installation ofregional flood concepts, etc.), and (3) risk management (e.g., flood forecasting,implementation of early-warning systems, and development of flood hazard andvulnerability maps, etc.).

This network will be developed as a flood warning network for assisting in real-time emergency services and will connected to a power database to effectivelyoptimise the flood management over the other existing methods by:

1) Providing access through a multiple-level web-based interface to a wide rangeof data types collected at investigated region in real-time. The user interfaceincludes a module for computer simulation of different flood scenarios, a tool formanaging simulation results, communication tools, etc. This, for example, willhelp the efficiency and optimize the stream gages locations and their operationsfor flood predictions (including early warning and cost-benefit analysis).

2) Combining the observational data with innovative data analysis to improveforecasting and risk assessment and analysis and providing a clear physical rep-resentation of the processes involved that can define risk zones and emergencyscenarios (which is one of the main limitations in the existing model) For exam-ple, values of water depth, velocity and their combination, and the flood timeare visualized in a global map, can provide a useful tool for emergency manage-ment and for determining protective measures against floods. This will support:


human interface and allow the technical-user to interact with the current repre-sentation of the design, enhance the user’s understandings, and make it quickerfor information to be reached on time and react properly to the warnings, etc.

3) Developing advanced computational methods for collecting, processing, gener-ating the data necessary for the fast and accurate simulation of different floodsituations and the 3D visualization of the numerical results. This will give thesynthesized results of monitoring data from different sources, models, data anal-ysis, etc. This will support evaluating the effect of alternative response scenariosby optimising the information overload (i.e., how to filter information and stillget the right information to the right people at the right time). Also, this willassist in establishing the social, economic and environmental goals for managingfloods.

4) Developing a flood warning network for assisting in real-time emergencyservices.

Metaheuristics can successfully handle a mix of continuous and discrete parametersas well as selecting individual components from database. This network will be con-nected to a database that combine environmental and geophysical data from earthobservation, satellite positioning systems, in-situ sensors and geo-referenced infor-mation with advanced computer simulation and graphical visualisation methods.This database will provide the following internet-based services: quickly locate andensure data availability where and when needed; the detailed description of contentsand limitations of the data; and present the data in different formats (maps, graphs,pictures, videos, etc.). In addition, this database will be designed to be searchable bydata type, data holder/owner, location, etc, and will be used in three modes: planningand design for flood protection; real-time flood emergency; and flood recovery.

As described above, flood protection is becoming more and more important. Inorder to be effectively prepared for floods, interdisciplinary and precautionary mea-sures with regard to water management, risk assessment, spatial planning, and landmanagement are necessary to reduce flood damage.

9 Conclusions and Future Work

The chapter presented a crucial step in disaster management by elucidatinghow dynamic optimisation and geo-information technologies could be efficientlyintroduced in the design process to support disaster risk reduction. Another innova-tive direction of the chapter shows how parallelisation and hybridisation of scientificresearch, technology development can effectively simplify handling data, minimizethe execution time, facilitate the design modelling using simulation and optimisa-tion process, handle robustness and simulate an appropriate behaviour of the designparameters in real-time. The aim of this chapter was to develop a new method-ology for effectively optimising the use of these technologies coupled with early


warning for increasing protection measures and reducing disaster damage. The mainconclusions and recommendations can be summarized as follows:

• Early warning systems are necessary for minimizing risks of global and localhazards by taking decisions in the proper time.

• Identifying and monitoring indicators and assessing environmental conditions areprerequisites for vulnerability assessments.

• Geo-information technologies provide important information source for decisionsupport and early warning systems.

• Scientific research and technology development are required.• Responsibilities of establishing early warning are shared among all parties

governments, communities and individuals through continuous capacity building.• Integration of risk management into development planning.• Commitment for, and belief, in the empowerment of poor through advanced

technology and continuous training and capacity building.

References

Brown C, Fingas M (2001) Upcoming satellites: potential applicability to oil spill remote sensing.In: Proceedings of the 24th Arctic and Marine Oil Spill Program (AMOP) technology seminar,Edmonton, Canada, June 12–14, pp 495–505

Cuny FC (1983) Disasters and development. Oxford University Press, New YorkDeb K (2001) Multi-objective optimisation & evolutionary algorithms. Chichester: WileyDorigo M, Gambarddella LM (1996) Ant Colony System: a cooperative learning approach to the

traveling salesman problem. IEEE Trans Syst Man Cybern B 26:29–41Egeland J (2006). Disaster risk reduction: a call for engagement. United Nations, New York,

October 10Elliott D (1996) Understanding GPS: principles & applications. Artech House, Boston, MAEuropean Environment Agency (2001) Late lessons from early warnings: the precautionary

principle 1886–2000. European Environment Agency, CopenhagenFedra K (1995) Distributed models and embedded GIS: strategies and case studies of integration.

In: Goodchild MF, Steyart LY, Parks BO, Johnston C, Maidment D, Crane M, Glendinning S(eds.) GIS and environmental modeling: progress and research issues. GIS World Books, FrontCollins, CO, pp 413–417

Fedra K (1994) GIS and environmental modeling. In: Goodchild MF, Parks BO, Steyaert LT (eds.)Environmental modeling with GIS. Oxford University Press, New York, pp 35–50

Glover F, Laguna M (1997) Tabu search. Kluwer Academic Publishers, Norwell, MAGoldberg DE (1989) Genetic algorithms in search, Optimization and machine learning. Addison-

Wesley Publishing Company, InC, Reading, MAHandmer J, Choong W (2006) Disaster resilience through local economic development. Aust J

Emerg Manage 21(4):8–15Jaszkiewicz A, Kominek P (2003) Genetic local search with distance preserving recombination

operator for a vehicle routing problem. Eur J Oper Res 151(2):352–364Kirkpatrick S, Gelatt CD, Vecchi PM (1983) Optimization by simulated annealing. Science

220:671–680Leick A (1995) GPS satellite surveying. Wiley, New YorkMasser I, Montoya L (2002) GIS in Urban Disaster Management. In City Development Strategies.

May 2002


Mileti D (1999) Disasters by design: a reassessment of natural hazards in the United States.Washington, DC: Joesph Henry Press

Oosterom PV, Zlatanova S, Fendel E (2005) Geo-information for disaster management.http://www.directionsmag.com/author.php?author_id=265. Accessed Feb 2005

Osman I, Kelly J (eds) (1992) Meta-heuristic: an overview. In Meta-heuristics: theory andapplications. Kluwer Academic Publishers, The Netherlands

Peng G, Leslie L, Shao Y (2002) Environmental modelling and prediction. Springer-Verlag, NewYork

Ramesh R, Eisenberg J, Schmitt T (eds) (2007) Improving disaster management: the role of IT inmitigation, preparedness, response, and recovery. National Academies Press, Washington, DC

Rekiek B, Delchambre A, Saleh H (2006) Handicapped person transportation: An application ofthe grouping genetic algorithm. J Eng Appl Artif Intell 19:511–520

Rene Vidal VV (1993) Applied simulated annealing. Springer, BerlinSaleh H (1996) Improvements to the GPSdemoUCL simulation software. MSc Dissertation,

Geomatic Engineering Department, University College London, UKSaleh H (2002) Ants can successfully design GPS Surveying Networks. GPS World 13(9):48–60Saleh H (2003) An artificial intelligent design for GPS surveying networks. J GPS Sol 7(2):

101–108Saleh H, Allaert G (2007) Monitoring network based optimisation and geo-information technology

for water pollution and water management in Flanders. In the Proceedings of Internationalconference: sustainable development & management of water in Palestine, Amman, Jordan,August 26, pp 234–246

Saleh H, Allaert G (2008) Real-life applications of disaster warning network based dynamicoptimisation and geo-information technology. In the proceedings of the 1st regional confer-ence on geoinformatics for disaster management and early warning systems, State of Kuwait,November, pp 175–189

Saleh H, Allaert G, De Sutter R, Kellens W, De Maeyer Ph, Vanneuville W (2008) Intelligentdecision support system based geo-information technology and spatial planning for sustainablewater management in Flanders. In: Feyen J, Shannon K, Neville M (eds) Water & urban devel-opment paradigms: towards an integration of engineering, design and management approaches.CRC Press, Taylor & Francis Group, London, pp 283–288

Saleh H, Dare P (2001) Effective heuristics for the GPS survey network of Malta: simulatedannealing and tabu search techniques. J Heuristics 7(6):533–549

Saleh H, Dare P (2002a) Heuristics for improved efficiency in the use of the GNSS for establishingpositioning networks. Marie Curie Fellowship Ann II:62–74

Saleh H, Dare P (2002b) Heuristic methods for designing GPS Surveying Network in the republicof Seychelles. Arabian J Sci Eng 26:73–93

Saleh H, Dare P (2003) Near-optimal design of Global Positioning System Networks using TabuSearch Technique. J Glob Optim 25:183–208

Saleh HA, Chelouah R (2003) The design of the Global Navigation Satellite Surveying Networksusing Genetic Algorithms. J Eng Appl Artif Intell 17/1:111–122

Turner B, Pidgeon N (1997) Man-made disasters, 2nd edn. Butterworth, LondonWattegama C (2007) ICT for disaster management, United Nations Development Programme –

Asia-Pacific Development Information Programme (UNDP-APDIP). http://www.apdip.netWisner B, Blaikie P, Cannon T, Davis I (2004) At risk: natural hazards, people’s vulnerability and

disasters, 2nd edn. Routledge, London

chapter 19 scientiﬁc research based optimisation …...chapter 19 scientiﬁc research based...

Documents